Photochemical nucleophile mapping: Identification of Tyr311 within the catalytic domain of rabbit muscle glyceraldehyde-3-phosphate dehydrogenase

Article abstract of DOI:10.1562/2006-02-28-RA-824

Photochemical mapping of nucleophiles in close proximity to the active site Cys149 of rabbit glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was demonstrated based on the nucleophilic aromatic photosubstitution reaction using two regioisomers of alkoxy-fluoro-nitro-substituted benzenes. Two photophores were covalently attached to the active site SH group of GAPDH and the protein was subjected to photolysis then to the cyanogen bromide cleavage reaction. The advantage of this method is the capability to chase labeled products by monitoring absorption at 380 nm because of the chromogenic property of photophore. HPLC separation identified a large labeled peptide fragment that was further digested by V8 protease for Edman sequence analysis. From the recent X-ray crystallography of rabbit GAPDH, Tyr311, His176, Ser238 and Lys183 are closely located to catalytic Cys149. Among these nucleophiles, Tyr311 was preferentially labeled with 2-fluoro-4-nitrophenoxy photophore and no label was identified with the isomeric 4-fluoro-2-nitrophenoxy photophore. The result clearly reflects the distance between Cys149 and nucleophiles to distinguish the nearest Tyr311. As photophores show great reactivity even with water under neutral conditions, the distance between nucleophiles and photophores is important for photo-induced nucleophilic aromatic substitution. The method will provide a useful technique to survey nucleophiles within the catalytic domain.

Full text of DOI:10.1562/2006-02-28-RA-824

Photochemistry and Photobiology, 2007, 83: 213–217
Photochemical Nucleophile Mapping: Identification of Tyr311 Within
the Catalytic Domain of Rabbit Muscle Glyceraldehyde-3-phosphate
Dehydrogenase†
Yasumaru Hatanaka*, Masaki Kaneda and Takenori Tomohiro
Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan
Received 28 February 2006; accepted 27 June 2006; published online 30 June 2006 DOI: 10.1562/2006-02-28-RA-824
functional group and irradiation produces very reactive species
to nonselectively insert into spatially proximate groups (2,3).
ABSTRACT
Photochemical mapping of nucleophiles in close proximity to the
active site Cys149 of rabbit glyceraldehyde-3-phosphate
dehydrogenase (GAPDH) was demonstrated based on the
nucleophilic aromatic photosubstitution reaction using two
regioisomers of alkoxy-fluoro-nitro-substituted benzenes. Two
photophores were covalently attached to the active site SH
group of GAPDH and the protein was subjected to photolysis
then to the cyanogen bromide cleavage reaction. The advantage
of this method is the capability to chase labeled products by
monitoring absorption at 380 nm because of the chromogenic
property of photophore. HPLC separation identified a large
labeled peptide fragment that was further digested by V8
protease for Edman sequence analysis. From the recent X-ray
crystallography of rabbit GAPDH, Tyr311, His176, Ser238 and
Lys183 are closely located to catalytic Cys149. Among these
nucleophiles, Tyr311 was preferentially labeled with 2-fluoro-
4-nitrophenoxy photophore and no label was identified with the
isomeric 4-fluoro-2-nitrophenoxy photophore. The result clearly
reflects the distance between Cys149 and nucleophiles to
distinguish the nearest Tyr311. As photophores show great
reactivity even with water under neutral conditions, the distance
between nucleophiles and photophores is important for photo-
induced nucleophilic aromatic substitution. The method will
provide a useful technique to survey nucleophiles within the
catalytic domain.
Ary azides, phenyl diazirines and benzophenones have been
frequently used for photo-affinity labeling. They usually
generate a nitrene, carbene or an excited carbonyl, respect-
ively, as a highly reactive intermediate, which allows immedi-
ate cross-linking with almost any functional group around the
binding site often encountered with analytical difficulties as a
result of the multi-site labeling.
Taking the advantage of photo-affinity labeling, the
development of a nucleophile-selective photophore should
solve the scrambling problem of labeled sites because of the
nonselective reactivity of common photophores. To survey
and profile nucleophiles in an unknown active center, or to
study the function of target nucleophiles by masking or
introducing another functional group, we developed a labe-
ling method using photolabile fluoronitrophenoxy group (4).
2-fluoro-4-nitroanisole was reported to react even with water
under neutral conditions based on the nucleophilic aromatic
photosubstitution reaction mechanism (5). The photochemi-
cal nature of 2-fluoro-4-nitroanisole satisfies the requirements
of photophores, which are chemically stable before photoac-
tivation, rapidly photolysed and highly reactive to almost all
nucleophiles. The nitrophenoxy group also provides the
capability to chase labeled products by monitoring absorp-
tion at 380 nm because of the chromogenic property of
photophores (6,7). In this study, we evaluated fluoronitro-
phenoxy photophores by probing the active site nucleophiles
surrounding the catalytic Cys149 residue of rabbit-muscle
GAPDH.
INTRODUCTION
In addition to the active center, enzymes usually require many
nucleophilic residues for the specific manipulation of their own
substrates. Numerous studies have been reported to elucidate
the relationship between activity and nucleophilic residues at
the active site. Affinity labeling is one of the most powerful
techniques for probing active site nucleophiles based on the
selective reactivity of properly designed reactive probes (1);
however, the reactivity of affinity probes is unmasked to result
in nonspecific labeling of protein surface nucleophiles. On the
other hand, photo-affinity labeling uses a chemically inert
MATERIALS AND METHODS
Materials and instruments. GAPDH was purchased from Roche
Molecular Biochemicals, Japan. Kieselgel 60 (70–230 mesh, Merck)
was used for column chromatography. All chemicals were of
analytical grade and were used without further purification. Melting
points were measured on Yamato RD-41 and uncorrected. Amino
acid profile was obtained on Hitachi 835, and sequence analyses
were performed by Perkin-Elmer Model 477A Protein Sequencer and
Model 120A PTH Analyzer. 1H NMR spectra were recorded on
JEOL JNM FX-200 spectrometer. UV–vis absorption spectra were
obtained on Hitachi M330 and photolysis were performed by UVP
Black-Ray MB-100A. Mass spectra (MS) and high-resolution mass
spectra (HRMS) were measured on JEOL JMS-300 and JEOL JMS-
DX3000, respectively.
†This invited paper is part of the Symposium-in-Print: Photobiology in Asia.
*Corresponding author email: yasu@pha.u-toyama.ac.jp (Yasumaru Hatanaka)
ꢀ 2007 American Society for Photobiology 0031-8655/07
213

214 Yasumaru Hatanaka et al.
Synthesis of 2-(fluoronitrophenoxy)ethyl methanethiosulfonate
(1). 2-(2-Fluoro-4-nitrophenoxy)ethyl chloride (2.20 g, 10 mmol) and
sodium methanethiosulfonate (4.02 g, 30 mmol) were dissolved in
10 mL DMF and the mixture was stirred at 120ꢁC for 1.5 h. To the
mixture was added 100 mL benzene, and the precipitate was removed by
filtration. After removal of solvent, the product was purified by
chromatography on silica gel eluted with benzene. Recrystallization
from benzene resulted 2-(2-fluoro-4-nitrophenoxy)ethyl methanethio-
sulfonate (1a) as colorless leaflets, 1.78 g (60%). M.p. 122–123ꢁC. IR
(nujor)1600 cm)1. 1H NMR(CDCl3) d 7.7–8.2(m, 2H), 7.0–7.1(m, 1H),
4.47 (t, 2H, J = 6 Hz) and 3.44 (s, 3H). MS m ⁄ z 295 (M+). Anal. Calcd.
for C9H10FNO5S2:C, 36.61;H, 3.41; F, 4.74; N, 6.43; S, 21.71. Found:C,
36.56; H, 3.37; F, 4.74; N, 6.46; S, 21.91. 2-(4-fluoro-2-nitrophen-
oxy)ethyl methanethiosulfonate (1b) was similarly prepared from
2-(4-fluoro-2-nitrophenoxy)ethyl chloride (3.14 g, 20 mmol). Recrys-
tallization from ether yielded colorless leaflets, 3.8 g (65%). M.p. 45–
46ꢁC. IR (nujor) 1590 cm)1. 1H NMR (CDCl3) d 7.61 (dd, 1H, J = 3,
8 Hz), 7.30 (ddd, 1H, J = 3, 8, 9 Hz), 7.12 (dd, 1H, J = 4, 9 Hz), 4.43
(t, 2H, J = 6 Hz), 3.57 (t, 2H, J = 6 Hz) and 3.40 (s, 3H). MS m ⁄ z 295
(M+). Anal. Calcd. for C9H10FNO5S2: C, 36.61; H, 3.41; F, 4.74; N, 6.43;
S, 21.71. Found: C, 36.54; H, 3.50; F, 4.79, N, 6.49; S, 21.77.
Preparation of photoreactive GAPDH. Before modification, rabbit-
muscle GAPDH was treated with charcoal (Sigma–Aldrich, USA) to
remove NAD+ according to the literature (11). To 0.1 M Tris buffer
(pH 8.5, 21 mL) containing the resulting apo-enzyme was added
2-(2-fluoro-4-nitrophenoxy)ethyl methanethiosulfonate (1a) or 2-(4-
fluoro-2-nitrophenoxy)ethyl methanethiosulfonate (1b) (660 lM in
DMSO), and the mixture was stirred at 4ꢁC for 30 min in the dark
to give GAPDH-a and GAPDH-b, respectively. The reaction mixture
was subjected to a gel filtration on Sephadex G-25 (1.15 · 30 cm,
pyrophosphate pH 8.5, 0.1 mM EDTA, 4ꢁC) to remove the excess
reagents. The GAPDH fractions were collected by monitoring of
absorption at 280 nm and used for the photolabeling experiments.
Photolysis of modified GAPDH and determination of labeling
efficiency. A solution of GAPDH-a or GAPDH-b in 10 mM
pyrophosphate buffer (pH 8.5) containing 0.1 mM EDTA was
irradiated with a 100 W black-ray lamp for 2 h at 0ꢁC. To determine
the efficiency of photo-affinity labeling, 50 mM pyrophosphate
buffer (pH 8.5, 0.8 mL) and 2-mercaptomethanol (1 lL) were added
to a photolabeled GAPDH sample solution (0.2 mL) followed by
incubation for 12 h at 20ꢁC. After ultrafiltration through Centricon
10 (Amicon) at 5000 g for 60 min, UV–vis spectra of the filtrates
were measured and the efficiencies were calculated from the
absorptions at 400 nm for GAPDH-a and 420 nm for GAPDH-b,
respectively.
Treatment of photoproducts with cyanogen bromide. After photoly-
sis of GAPDH-a, the sample solution (40 mL) containing urea
(19.6 g, 8 M), EDTA (84 mg, 0.2%) and 2-mercaptomethanol
(64 lL, 0.9 mmol) was added. The solution was adjusted to pH 8.2
by sodium hydroxide, and incubated at 20ꢁC for 4 h. After
incubation with iodoacetic acid (166 mg, 0.8 mmol) at 20ꢁC for
1 h, the mixture was dialyzed in water and then concentrated
in vacuo. The sample was dissolved in 70% formic acid (27 mL)
followed by the addition of cyanogen bromide (95 mg, 0.9 mmol).
The mixture was incubated at 20ꢁC for 24 h under argon. To the
solution was added water (30 mL) and concentrated to 2 mL
in vacuo. This procedure was repeated thrice. The solution was
centrifuged and the pellet was dissolved in formic acid (1 mL) and
combined with the supernatant. The sample solution was loaded on a
Chemcosorb 7C18 column (10 · 300 mm, Chemco) which was pre-
equilibrated with 26% acetonitrile containing 0.1% trifluoroacetic
acid and developed with a linear gradient of acetonitrile consisting of
26–58% over 60 min at flow rate of 2 mL ⁄ min. Fractions were
monitored by measuring the absorbance at 215 and 380 nm. The
main peak fractions were pooled and the solvents were evaporated to
1 mL in vacuo. The sample was chromatographed on l Bondapak
C18 (10 · 300 mm, Waters) which had been pre-equilibrated with
0.1% trifluoroacetic acid containing 18% acetonitrile-2-propanol (3 :
7, v ⁄ v), followed by a linear gradient of 18–82% acetonitrile-2-
propanol (3 : 7, v ⁄ v) over 30 min at a flow rate of 1 mL ⁄ min.
Fractions were monitored by measuring the absorbance at 215 and
380 nm. The main peak fractions were pooled and the solvents were
evaporated in vacuo. Treatment of GAPDH-b and isolation of the
labeled peptides were similarly performed.
V8 protease digestion of photoproducts. The isolated peptides after
cyanogen bromide treatment were digested with V8 protease
(1 mg ⁄ mL, 200 lL, 7.2 nmol) in 100 mM sodium phosphate buffer
(pH 7.8, 20 mL) at 37ꢁC for 12 h followed by continuous incubation
for 36 h after addition of the same amount of V8 protease. The
solution was concentrated to 0.5 mL in vacuo. The solution was
centrifuged and the pellet was dissolved in formic acid (0.3 mL) and
combined with the supernatant. The sample was chromatographed on
micron Bondapak C18 (10 · 300 mm, Waters) which had been pre-
equilibrated with 18% 2-propanol-acetonitrile (7 : 3, v ⁄ v) containing
0.1% trifluoroacetic acid, followed by a linear gradient of 18–50%
2-propanol-acetonitrile (7 : 3, v ⁄ v) over 30 min at a flow rate of
1 mL ⁄ min. Fractions were monitored by measuring the absorbance at
215 and 380 nm. The main peak fractions were pooled and the solvents
were evaporated to 1 mL in vacuo. The solution was centrifuged and
the pellet was dissolved in formic acid (0.5 mL) and combined with the
supernatant. The sample solution was loaded on a Chemcosorb 7C18
column (4.6 · 250 mm, Chemco) which was pre-equilibrated with 20%
acetonitrile containing 0.1% trifluoroacetic acid and developed with
20% acetonitrile containing 0.1% trifluoroacetic acid at flow rate of
1 mL ⁄ min. After removal of the solvent, the residue was dissolved in
70% formic acid for amino acid sequence analysis. Treatment of
GAPDH-b and purification of the labeled peptides were similarly
performed.
Photoreaction of fluoronitroanisoles and N-acetyllysinamide.
2-Fluoro-4-nitroanisole (2a, 171 mg, 1 mmol) and N-acetyllysinamide
(187 mg, 1 mmol) was dissolved in a mixed solvent (20 mL) of 50%
(v ⁄ v) acetonitrile and 0.1 M aqueous sodium hydrogen carbonate. The
solution was irradiated with a 100 W black-ray lamp for 5 h under
nitrogen. After removal of the solvent, the product was chromato-
graphed on silica gel (chloroform : methanol = 20 : 1) followed by
recrystallization from ethyl acetate to give N-acetyl-N¢-(2-methoxy-
5-nitrophenyl)lysinamide (3a) as yellow needles (12 mg, 4%). M.p.
184–185ꢁC. IR (nujor) 3320, 3200, 1700, 1680 cm)1. 1H NMR (CDCl3)
d 7.62 (dd, 1H, J = 3, 9 Hz), 7.34 (d, 1H, J = 3 Hz), 6.74 (d, 1H,
J = 9 Hz), 6.07 (br s, 2H), 5.39 (br s, 1H), 4.48 (dt, 2H, J = 7, 7 Hz),
4.39 (br s, 1H), 3.94 (s, 3H), 3.21 (dt, 2H, J = 7, 7 Hz), 2.03 (s, 3H)
and 1.4–2.0 (m, 6H). MS m ⁄ z 388 (M+). HRMS m ⁄ z calcd. for
C15H22N4O5: 388.1592. Found: 338.1602. 2-Methoxy-5-nitrophenol
(21 mg, 12%) was also given in the reaction and 2a (137 mg, 80%) was
recovered. The photoreaction of 4-fluoro-2-nitroanisole (2b) for 10 h
gave N-acetyl-N¢-(4-methoxy-3-nitrophenyl)lysinamide (3b) as orange
needles after recrystallization from ethyl acetate (26 mg, 8%). M.p.
132–133ꢁC. IR (nujor) 3350, 3200, 1700, 1675 cm)1. 1H NMR (CDCl3)
d 7.05 (d, 1H, J = 3 Hz), 6.94 (d, 1H, J = 9 Hz), 6.78 (dd, 1H,
J = 3, 9 Hz), 6.09 (br s, 2H), 5.42 (br s, 1H), 4.51 (dt, 2H, J = 7,
7 Hz), 3.88 (s, 3H), 3.72 (br s, 1H), 3.12 (t, 2H, J = 7 Hz), 2.03 (s, 3H)
and 1.4–2.0 (m, 6H). MS m ⁄ z 388 (M+). HRMS m ⁄ z Calcd. for
C15H22N4O5: 388.1592. Found: 338.1602. 4-Methoxy-3-nitrophenol
(31 mg, 18%) was also given in the reaction and 2b (116 mg, 67%) was
recovered.
RESULTS AND DISCUSSION
Preparation of photoreactive GAPDH
GAPDH is a homo-tetramer protein and each subunit has an
active site containing an un-oxidized sulfhydryl group of
Cys149 used for catalytic reaction. It catalyzes the oxidation of
glyceraldehyde-3-phosphate to form 1,3-diphosphoglycerate
with the concomitant reduction of b-nicotinamide adenine
dinucleotide (NAD+) (8–10). The chemical labeling of the
rabbit-muscle GAPDH active site with thermally reactive
probes suggested that the ꢀ-amino group of Lys183 could be
near to catalytic SH because it was acetylated through the
migration of acetyl Cys149 (15) from S to N atom and this
observation was confirmed by a cross-linking reagent (16). The
X-ray structure of rabbit-muscle GAPDH was recently repor-
ted (11), and the catalytic center was revealed as the first
detailed model of mammalian GAPDH. There are several

Photochemistry and photobiology, 2007, 83 215
˚
nucleophiles existing within 6 A from the active site Cys149 of
one subunit of GAPDH-a and -b, respectively. The concen-
tration of GAPDH protein itself was determined from the
absorption at 278 nm.
rabbit-muscle GAPDH, such as His176 and Tyr311, whereas
the distance to Lys183 is about 3 times longer than these
nucleophiles. For the photochemical nucleophile mapping of
rabbit-muscle GAPDH, fluoronitrophenyl derivatives having a
SH-reactive methanethiosulfonyl group (1a and 1b, Fig. 1)
were designed for attaching photophores to the catalytic
Cys149 via disulfide linkage (12). Because nucleophilic aroma-
tic photosubstitution takes place at the fluorinated position of
the benzene ring, ortho- and para-derivatives were prepared to
compare labeling efficiency and ability depending on the
distance from Cys149 to nucleophiles.
As bound NAD+ inhibits the modification of the Cys149 SH
group (13), the cofactor was removed before the incorporation
of photoreactive moiety. The reaction of apo-GAPDH and
compound 1a or 1b efficiently proceeded at 4ꢁC to give
photoreactive GAPDH proteins (GAPDH-a and GAPDH-b,
respectively). The reaction was monitored by measuring the
enzyme activity of GAPDH according to the literature (9).
GAPDH activities after 30 min were 0.08 and 0.03% of the
original activity by the reaction of compounds 1a and 1b,
respectively. The electronic absorption spectra of photoreac-
tive GAPDH-a and -b are shown in Fig. 2 (spectra at 0 min).
The modification yield was calculated based on absorption
because of the fluoronitrophenoxy group by comparison with
the values of glutathione derivatives of compounds 1a and 1b
[kmax 314 nm (ꢀ 11500) and 343 nm (ꢀ 2630), respectively] as
1.01 and 0.99 molar ratios of fluoronitrophenoxy groups per
Photoreactions of modified GAPDH
Photolysis was performed at 0ꢁC with a 100 W long-wavelength
UV lamp until specific absorption as a result of photoreactive
moieties, fluoronitrophenoxy groups, disappeared. Figure 2
shows UV–vis spectra of photoreactive GAPDHs and the time
course of degradation of photoreactive groups by irradiation.
Absorption around 310–320 nm for GAPDH-a and 320–
340 nm for GAPDH-b clearly decreased with the irradiation
time. These results show that photolysis was completed within
120 min. The electrophilic intermediates produced during
photolysis should mainly react with spatially proximate nucle-
ophiles at the active site of GAPDH to form covalent bonds
with amino acid residues. However, photoreactive fluoronitro-
phenoxy groups also reacted with the surrounding water
molecules and were converted to nitrophenol derivatives. After
treatment with 2-mercaptoethanol to cleave the disulfide bond,
nitrophenol derivatives should be separated and isolated by
ultra-filtration from the protein. The electronic absorption
spectra of the filtrates were measured to determine the amount
of ‘unlabeled’ nitrophenol derivatives and the labeling yields
were then calculated based on the ꢀ values of 2-hydroxy-
4-nitroanisol (4a for GAPDH-a) and 4-hydroxy-2-nitroanisol
(4b for GAPDH-b) to give 9% and 12%, respectively.
Identification of the photolabeled sites
The photolabeled proteins were then subjected to amino acid
analysis of the labeled fragments and determination process of
the labeling sites.
Firstly, photolabeled GAPDH-a and GAPDH-b were
cleaved with cyanogen bromide after reductive carboxymethy-
lation. The resultant fragments were purified by HPLC by
monitoring electronic absorption at 215 and 380 nm. As a
nitrophenoxyl group shows absorption around those wave-
lengths, the labeled peptide tethering it can be pursued by
spectral analysis. A broad band eluted around 28–32 min was
isolated (Fig. 3a) and subjected to amino acid analysis. The
chromatograms showed that the isolated fragments contained
many acidic residues and a large amount of unlabeled peptides.
The fragments were then digested by V8 protease and the
proteolytic fragments were separated by HPLC followed by
isolation of the fragment eluted at 28 min (indicated by arrow
in Fig. 3b) that showed absorption at 380 nm. A similar result
was obtained for GAPDH-b. The result of Edman sequence
analysis for GAPDH-a fragments is summarized in Table 1. It
showed that the photolabeled peptide was equivalent to
residues 303–314. The amino acid residue 310 (cycle 8) was
analyzed as alanine instead of tryptophan. This region is well
conserved from bacteria to mammals and the residue at this
position was already identified as tryptophan (11, see SUP-
PLEMENTAL MATERIALS). As it was reported that
tryptophan residue was oxidized by cyanogen bromide (14),
a similar reaction should occur in this case. Furthermore, the
product of cycle 9 was not detected at the expected retention
time of PTH-Tyr derivative on HPLC (see SUPPLEMENTAL
MATERIALS). The result suggested that Tyr311 in GAPDH
Figure 1. 2-(Fluoronitrophenoxy)ethyl methanethiosulfonate com-
pounds as photoreactive reagents with a disulfide linkage.
Figure 2. UV–vis spectral changes (240–500 nm) of photoreactive
GAPDHs by irradiation are shown. (a) GAPDH-a modified by
2-(2-fluoro-4-nitrophenoxy)ethyl methanethiosulfonate. (b) GAPDH-
b modified by 2-(4-fluoro-2-nitrophenoxy)ethyl methanethiosulfonate.

216 Yasumaru Hatanaka et al.
Figure 3. Chromatogram profiles of proteolytic products of photolabeled GAPDH-a. The eluates were chased by monitoring absorptions at 215
and 380 nm. (a) Chromatogram of degradation sample of photolabeled GAPDH-a with cyanogen bromide. The fraction indicated by the bar was
collected and subjected to amino acid analysis. (b) Chromatogram of the digestion sample of photo-labeled GAPDH-a with V8 protease. The
indicated fraction (arrow) was collected for Edman sequence analysis.
Table 1. Determination of amino acid sequence of the labeled peptides
for position 303–314
PTH amino acids (pmol)
Cycle
V8 fragment-a
V8 fragment-b
1
2
3
4
5
6
7
8
9
His (220)
Phe (1144)
Val (1249)
Lys (560)
Leu (1024)
Ilu (1023)
Ser (105)
Ala (144)
Xaa (–)a
His (44)
Phe (752)
Val (489)
Lys (212)
Leu (462)
Ilu (397)
Ser (24)
Ala (159)
Tyr (71)
Asp (41)
Asn (35)
Glu (20)
Figure 4. Photoreaction of 2-fluoro-4-nitroanisole with N-acetyllysin-
amide.
examined in dimethylformamide or higher basic solution
(pH 12), similar results were obtained. On the other hand,
the photoreaction using N-acetyllysinamide was also studied
(Fig. 4) because it has an amino group that is usually
considered as a more reactive nucleophile than phenol group.
In addition, it has been suggested that S to N migration of the
acetyl group from Cys149 to Lys183 could proceed via an
intermediary group (15,16). Lysine adducts were detected and
isolated in 4% and 8% yields for compounds 3a and 3b,
respectively. These results indicate that nucleophilic substitu-
tion of the photophores depends on the basicity of nucleo-
philes, and the competitive quenching reaction with water is
sufficiently slow even under basic conditions. While the labeled
peptide was unfortunately not identified for GAPDH-b, only
Tyr311 was clearly labeled in the photolysis of GAPDH-a. In
terms of the reactivity of the fluoronitrophenoxy group against
tyrosine, it was suggested that 2-fluoro-4-nitrophenoxy groups
should exist spatially proximate to Tyr311. Actually, Fig. 5
10
11
12
Asp (189)
Asn (262)
Glu (30)
aNot detected.
was labeled by 2-fluoro-4-nitrophenoxy group upon photoly-
sis. In the case of GAPDH-b, the amount of PTH derivatives
was rather small such as Ser309, Asp312, Asn313 and maybe
Tyr311 (Table 1). However, the labeled amino acid residues
were not clearly detected in this study, although the isolated
peptide showed absorption at 380 nm.
As Tyr311 was labeled in this study, the photochemical
reactivity of fluoronitroanisole derivatives, 2-fluoro-4-nitroan-
isole (2a) and 4-fluoro-2-nitroanisole (2b), was investigated
with N-acetyltyrosinamide for confirmation (see SUPPLE-
MENTAL MATERIALS). The photoreactions of fluoronit-
roanisoles and N-acetyltyrosinamide were carried out in 0.1 M
carbonate buffer (pH 8.0) - acetonitrile (1: 1) under an inert
atmosphere. Both reactions slowly proceeded, resulting in the
recovery of more than 75% of starting material after 10 h
irradiation. The products were mainly corresponding phenol
derivatives derived from the reaction with water, and no
tyrosine adducts were detected. Although the reactions were
˚
shows that the distance between Cys149 and Tyr311 is 4.9 A,
˚
estimated from the X-ray structure (11), comparable to ca 6 A
of the distance between Cys149 and the ortho-position (reac-
tion site) of the benzene ring. In the case of the 4-fluoro-
2-nitrophenoxy group, substitution at the para-position of the
benzene ring by tyrosine residue is supposed to be sterically
and conformationally difficult because of the rather long
˚
distance between the para-position and Cys149 (ca 8 A) as well

Photochemistry and photobiology, 2007, 83 217
Acknowledgements—This work was supported in part by CLUSTER
(Cooperative Link of Unique Science and Technology for Economy
Revitalization) and the Fugaku Trust for Medicinal Research. We
thank Dr. Eiichi Yoshida for his technical assistance with the
photolabeling experiments of GAPDH.
SUPPLEMENTAL MATERIALS
Figures S1, S2 and Table S1 can be found at www.blackwell-
synergy.com
REFERENCES
1. Colman, F. R. (1997) Chemical arrows for enzyme targets. FASEB
J. 11, 217–226.
2. Hatanaka, Y. and Y. Sadakane (2002) Photoaffinity labeling in
drug discovery and developments: Chemical gateway for entering
proteomic frontier. Curr. Top. Med. Chem. 2, 271–288.
3. Tomohiro, T., M. Hashimoto and Y. Hatanaka (2005) Cross-
linking chemistry and biology: Development of multifunctional
photoaffinity probes. Chem. Record 5, 385–395.
4. Brasem, P., J. G. Lammers, J. Cornelisse, J. Lugtenburg and E.
Hainga (1972) Photoreactions of aromatic compounds XXIV
nucleophilic photosubstitution of fluorobenzene derivatives.
Tetrahedron Lett. 8, 685–686.
5. Cornelisse, J. and E. Havinga (1975) Photosubstitution reaction of
aromatic compounds. Chem. Rev. 75, 353–388.
6. Hatanaka, Y., E. Yoshida, H. Nakayama and Y. Kanaoka (1989)
Chromogenic diazirine: A new spectrophotometric approach for
photoaffinity labeling. Bioorg. Chem. 17, 482–485.
7. Hatanaka, Y., M. Hashimoto, H. Kurihara, H. Nakayama and Y.
Kanaoka (1994) A novel family of aromatic diazirines for pho-
toaffinity labeling. J. Org. Chem. 59, 383–387.
Figure 5. Relationship between the oxidized active site of Cys149 and
surrounding amino acid residues as well as the adenine part of the
cofactor NAD+ (PDB ID: 1J0X (11)).
as the rigidity of photophores. These results obtained by
photolabeling obviously reflected the three-dimensional align-
ment of nucleophiles within the active site. Although the
amino group of lysine showed much higher reactivity, Lys
residue did not react with photophores in this experiment. The
distances to Lys183 and Lys306 from Cys149 are 17.8 and
8. Tanner, J. J., M. R. Hecht and L. K. Krause (1996) Determinants
of enzyme thermostability observed in the molecular structure of
thermus aquaticus D-glyceraldehyde-3-phosphate dehydrogenase
˚
at 2.5 A resolution. Biochemistry 35, 2597–2609.
˚
13.5 A, respectively, which are too far to make a bond.
Furthermore, there are other nucleophiles spatially proximate
9. Harris, I. J. and M. Waters (1976) Glyceraldehyde-3-phosphate
dehydrogenase. In The Enzyme, Vol. XIII part c 3rd ed. (Edited by
P. Boyer), pp. 1–49. Academic Press, New York.
10. Moras, D., N. K. Olsen, N. M. Sabesan, M. Buehner, C. C. Ford
and G. M. Rossmann (1975) Studies of asymmetry in the three-
dimensional structure of Lobster D-glyceraldehyde-3-phosphate
dehydrogenase. J. Biol. Chem. 250, 9137–9162.
˚
to Cys149, including His176 (4.4 A), Ser238 (6.3 A), Try317
˚
˚
˚
(8.0 A), Cys153 (8.8 A), as well as two adjacent amino acid
residues, Ser148 and Thr150. In a previous report, the
2-fluoro-4-nitorophenoxy derivative of glutathione was suc-
cessfully identified as a histidine residue within the substrate-
binding site of S-transferase (17). No labeled products of these
residues, including His176, were detected at this time, although
His176 exists close to Cys149 as well as Tyr311.
11. Cowan-Jacob, S. W., M. Kaufmann, A. N. Anselmo, W. Stark
and M. G. Grutter (2003) Structure of rabbit-muscle glyceralde-
hyde-3-phosphate dehydrogenase. Acta Crystallogr.
Crystallogr. 59, 2218–2227.
D Biol.
12. Kaneda, M., Y. Sadakane and Y. Hatanaka (2003) A novel
approach for affinity-based screening of target specific ligands:
Application of photoreactive D-glyceraldehyde-3-phosphate
dehydrogenase. Bioconjugate Chem. 14, 849–852.
CONCLUSION
13. Trentham, D. R. (1968) Aspect of the chemistry of D-glyceralde-
hyde-3-phosphate dehydrogenase. Biochem. J. 109, 603–612.
14. Boulware, D. W., P. D. Goldsworthy, F. A. Nardella and M.
Mannik (1985) Cyanogen bromide cleaves Fc fragments of pooled
human IgG at both methionine and tryptophan residues. Mol.
Immunol. 12, 1317–1322.
15. Mathew, E., B. P. Meriwether and J. H. Park (1967) The
enzymatic significance of A-acetylation and N-acetylation of
3-phosphoglyceraldehyde dehydrogenase. J. Biol. Chem. 242,
5024–5033.
16. Shaltiel, S. and M. T. Finkelstein (1971) Introduction of an
intermolecular crosslink at the active site of glyceraldehyde
3-phosphate dehydrogenase. Biochem. Biophys. Res. 44, 484–
490.
17. Nishihira, J., M. Sakai, S. Nishi and Y. Hatanaka (1995) Identi-
fication of the electrophilic substrate-binding site of glutathione
S-transferase P by photoaffinity labeling. Eur. J. Biochem. 232,
106–110.
We presented an example of nucleophile mapping based on a
nucleophilic aromatic photosubstitution reaction and demon-
strated that Tyr311 of rabbit-muscle GAPDH was mainly
photolabeled by a photoreactive and chromogenic 2-fluoro-
4-nitrophenoxy group attached to the active site of Cys149.
Substitution took place with Tyr residue although it showed
lower reactivity than Lys residue. In addition, labeling of
spatially proximate His176 did not occur in this study, since
the substitution reaction by His residue was already reported.
The result suggested that the distance and configuration of a
nucleophile to photophores must be of primary importance for
labeling.
In this study, a new method of photolabeling has been
developed for the mapping of nucleophiles located in the
proximity of fluoronitrophenoxy photophores.